Table 3.

Performance Models Used for Prediction of TMB Level According to Patient’s Receptor Status

Model	Accuracy ± SD (%)	Mean AUC	Hyperparameters
Logistic Regression	72.22 ± 6.4	0.83 ± 0.18	penalty = L2 and C= 1.0
Kernel SVM	71.73 ± 12.04	0.76 ± 0.17	C = 1, kernel is Gaussian, and gamma is 0.01.
K Nearest neighbor (KNN)	74.4 ± 9.76	0.81 ± 0.15	Algorithm: auto, leaf_size: 1, n_jobs: -1, n_neighbors: 5
Decision tree	74.11 ± 5.99	0.81 ± 0.25	Criterion: gini , max_depth: 4 , Number Of Components: 3
Random forest tree	71.73 ± 12.04	0.73 ± 0.17	max_depth: 5, min_samples_leaf: 1, n_estimators: 25